Windows 8 Pro BSOD help!

[chrisgt]

16gb Kingston hyperx genesis grey 1600mhz.

Bought 8mb from ocuk in dec 2011 then another 8gb in September 2012. Had all 16gb in my old pc running at 1600ghz with no problems. Memtest86 is running as we speak so will leave that until morning to see if any errors show up..then will try stick by stick

 

[chrisgt]

One mistake here is the mixing RAM. You shouldn't mix RAM even if there same Model. Thats why you buy in sets 4GB, 8GB and 16GB
See how the Memtest runs if error show then you will need to find witch RAM stick is damaged and replace it.

Even if the 2x matched sticks are in dimm slot 1 and 3, and the other two in dimm slots 2 and 4??

This has worked for months though...

 

[chrisgt]

I have tried mirroring the Kingston settings as shown on their site and the CPUZ SPD tab...800mhz 9-9-9-27 1.65v but that crashes. Even at BIOS optimised defaults, which is something like 667mhz 11-11-11-28 1.50v, it crashes on Blend...

This is infuriating! The RAM used to be fine...but what I don't get is if I leave the RAM alone and just try OCing the CPU it crashes...

 

[chrisgt]

MemTest ran for 12hours, no errors:

 

[chrisgt]

So I tried once again configuring the RAM at 800Mhz 9-9-9-27 1.65v which immediately caused a crashed in Prime95 Blend test, but I've been running the FPU test (top one- lots of CPU, little RAM) for about 15minutes without issue.

So this is still pointing to the RAM? Despite the fact MemTest found no errors? I have 4GB 1600MHZ Corsair XMS3 9-9-9-24 1.65v from another system which I will try out in this and report back.

 

[chrisgt]

With the 4GB Corsair in, using XMP Profile 1 in the BIOS (which is correct @ 9-9-9-24 1.65v) I bluescreen 10 seconds into a Blend test. Same error as before, different RAM.

So:
1) 16GB Kingston RAM is causing crashing when run at correct speeds (800Mhz) or even when under-clocked to 667Mhz
2) 4GB Corsair RAM is causing crashing when run at correct speeds (800Mhz). (http://www.corsair.com/cmx4gx3m2a1600c9.html)

What else might this be?

 

[chrisgt]

I would have said this is likely due to the Z77 chipset recommending DRAM Voltage is @ 1.5v, but my memory requires 1.65v to achieve the 800Mhz it should.

But even when it was at 667Mhz and 1.5v I got a BSOD playing PoE.

http://www.tomshardware.co.uk/forum/319559-12-memory
http://hardforum.com/showthread.php?t=1686478
http://teksyndicate.com/forum/motherboards/asus-sabertooth-z77-15v-vs-165v-memory/126897

Related to me:
http://www.mmo-champion.com/threads/1165784-ASUS-Sabertooth-Z77-1-65v-RAM

 

[chrisgt]

CPU or Motherboard Do you have spear thermal paste? Can you re-seat the CPU I would also check the Pins on the socket they is a chance you might have bent them.
The CPU has a Memory Controller so if this is damaged would explain why you keep crashing when stressing the RAM.

Mind you this is all just guess work right now.

I have already re-seated the CPU and re-fitted the CPU cooler (with new paste).

Not true Z68 also recommends 1.5v and my RAM is rated 1.65v but I run mine @1.575v

I have clocked my RAM back to 667mhz @ 1.5v and now Blend test runs (for 15 minutes so far). Which is a damn sight better than previously, where it would crash after 10 seconds. So it does appear to be the RAM voltage BUT I had the RAM at 667Mhz @ 1.5v before and still got 1 BSOD.

I will leave Prime95 Blend running for a few hours (need to pop out to run some errands) and report back. If that is stable, I might look at buying some new RAM (8GB which is rated @ 1600Mhz @ 1.5v)

EDIT: Blend is still going...hasn't crashed yet!!! So it looks as if its 1) definitely a memory issue and 2) definitely a voltage related issue with the memory...

 

[chrisgt]

Blend test still going...2+ hours without BSOD.

Looks to me like its definitely the DRAM Voltage causing the issue...

In which case I need to look for 8GB of RAM that runs at 1600Mhz @ 1.5v! Any recommendations??

More evidence that > 1.5v can cause instability: http://communities.intel.com/thread/30798

 

[chrisgt]

But the RAM I'm testing with now has to be run @ 667Mhz in order to keep it at 1.5v. I'd want 800Mhz @ 1.5v

 

[chrisgt]

I use http://www.overclockers.co.uk/showproduct.php?prodid=MY-363-CS&groupid=701&catid=8&subcat=1517 on my Z68 Asus board never had any issues.

Those look good - XMP 11-11-11-30 values at 1600MHz, 1.50V

Need as little heatsink as possible as the CM CPU Cooler is so fat.

 

[chrisgt]

Been testing this out more today...

Prime95 Blend test - 4 hours passed without error with settings:

Set-up 1: BIOS Optimized Defaults
i5 @ 3.4Ghz - Completely Stock (stock voltages (everything at auto in BIOS))
RAM @ 667Mhz 9-9-9-24 1.5v

So I'm pretty happy that so long as the RAM is under-clocked to 667Mhz and running at 1.5v, things will be pretty stable. I then tried the below set-up:

Set-up 2: Slight Multiplier OC
i5 @ 4.0Ghz (40x100BLCK with stock voltages)
RAM @ 667Mhz 9-9-9-24 1.5v

The above Blue Screened after about 30 seconds into the Prime95 Blend test. Can this still be attributed to the RAM, even though nothing RAM related changed between set-up 1 and set-up 2??? RAM was running at the exact same speeds/voltages for both set-ups...it was only the CPU multiplier changed between the two, yet set-up 2 insta-crashed...

 

[chrisgt]

OK if I leave the memory settings alone and focus just on the CPU, what would you recommend for a modest 4.0Ghz?

100BLCK X 40
Voltage Response to Fast
CPU Vcore @ ???

 

[chrisgt]

Set-up 1: BIOS Optimized Defaults
i5 @ 3.4Ghz - Completely Stock (stock voltages (everything at auto in BIOS))
RAM @ 667Mhz 9-9-9-24 1.5v

OH MY ******* GOD! That set-up, which did 4 hours in Prime without an error, that I've been using all day to play PoE and Blops 2, just gave me the same f'ing error WHEA_UNCORRECTABLE_ERROR about 1 minute ago...

So I don't even think it is memory voltage related anymore...I'm getting fed up with this!!!!!!!!!!!!!!!! What else could it possibly be??

 

[chrisgt]

I would run your ram at 1.55v for 1600mhz,you can set system performance enhance to standard if your on a gigabyte board,and try looser timings of 9-9-9-28-2t

I'm running it at 1.5v @ 1333Mhz just to keep the PC from crashing every few minutes. I will try going back to 1600mhz with a reduced voltage of 1.55v, however the manufacturer recommends 1.65v for that speed, so I don't think it will be stable!

Whea error is not enough CPU voltage,

Edit,you can be 100% stable with CPU stress but soon as you start gaming it will throw up whea errors so you have to up CPU v some more

But what I don't understand is how the CPU can not be getting enough voltage; it is at factory settings?? The lowest I'm seeing in HWMonitor is 1.092 and the highest is 1.260. Surely the chip wouldn't ship with such an issue?

What would you suggest I up the Vcore to?

 

[chrisgt]

Hi Chris, seems you are not having much luck at the moment. Can`t offer any info that will help solve your issues, other than suggest you open a thread in the link below. No disrespect to shankly or wazza who have been very helpful but the Overclocking forum would be a more informative place for help then a software forum.
Any way, I wish you good luck and hope you get things sorted sooner rather than later.

http://forums.overclockers.co.uk/forumdisplay.php?f=149

Thanks Idleman. I have posted in the Ask 8 Pack sub-forum, is that correct?